Distribution and Diagnostic Value of Pancreatic Enzymes in Dialysis Patients

Purpose: As serum levels of pancreatic enzyme are elevated in dial sis patients, clinicians ma encounter difficulty in diagnosing pancreatitis among dialysis patients with abdominal pain. In this work, we analyze the distrtbution of pancreatic enzyme levels in dialysis patients and compared different cut-off values in establishing diagnosis of acute pancreatitis in dialysis population. Patients and Methods: Serum levels of amylase and lipase were measured in 39 peritoneal dialysis (PD) patients, 76 hemodialysis (HD) patients and 266 people with normal renal function. Another 21 dialysis patients with pancreatitis cross-referenced from tile hospital computer database were also included in the analysis. Distribution of pancreatic enzymes between different groups was analyzed, and sensitivity/specificity of different cut-off values was compared using SPSS software. Results: Mean serum amylase and lipase level were higher in HD as well as PD patients thou in the normal controls (p＜0.05). In total, 67 of tile 115 (58%) asymptomatic dialysis patients had an anylase level higher than the upper normal limit with 2 (2%) was higher than 3 times tile normal. Eighty-five of 115 (74%) patients had a lipase level higher than the normal with 17 (15%) higher than 3 times the normal. The highest level of amylase was 5.3 times and lipase was 6.7 times the upper normal limit. When contracting pancreatitis, dialysis patients had further rise in amylase and lipase levels (p＜0.05 for both). However, significant overlapping in pancreatic enzyme levels existed between those with and without pancreatitis. ROC curve analysis revealed that the cut-off value providing the best balance between sensitivity and specificity was 247 U/L for amylase (0.714/0.839), and 631 U/L for lipase (0.905/0.902). When using the traditional 3 times the normal as the cut-off value, the sensitivity/specificity of amylase criterion of 420 U/L was 0.333/0.982 while that of lipase criterion of 570 U/L was 0.952/0.857. Alternatively, using 7 tunes rile upper normal limit as the cut-oft value provides a specificity of 1.0 for both amylase amid lipase criteria. Conclusion: Asymptomatic dialysis patients might have highly elevated pancreatic enzyme levels which overlap severely with patients with pancreatitis. Thus, diagnosis of acute pancreatitis should be made with care. A cut-off value of 3 times the upper normal limit provided an acceptable balance between sensitivity and specificity criteria for lipase hut not for amylase. A cut-off value of 247 U/L for amylase and 621 U/L for lipase is a the preferred criterion in our hospital. However, only the application of 7 times the upper normal limit as a the cut-off value could guarantee correct diagnosis of pancreatitis in dialysis population.